/atm/cam/sst/sst_HadOIBl_bc_48x96_1850_2012_c130411.nc

[cnrwil004@myuct_ac_za]

Hi all

We are trying to get an "eul" dycore F compset case (preferably at relatively low resolution) of CESM1.2.0 running on a local cluster. We have successfully managed to run a F compset case with finite volume dycore at 4x5 degree resolution (f45_f45). However, upon trying to run a case created with "./create_newcase -case ${CASENAME} -res T31_T31 -compset F -mach ${MACHNAME}", which successfully compiled, the following is obtained in the cesm.log.$LID file:

> " .
.
.
Remapping the ITD is not allowed for ncat=1
Using the delta function ITD option instead
CalcWorkPerBlock: Total blocks: 72 Ice blocks: 72 IceFree blocks: 0 Land blocks: 0
(shr_sys_abort) ERROR: ice_open_nc: Cannot open "$DIN_LOC_ROOT/DataDir/share/domains/UNSET"

From other posts in this forum I gathered that this might be an inherent problem with this dycore, compset, resolution combination. Could someone perhaps confirm that this is the case? Would this be considered a bug?

Following a suggestion from a different post in this forum, we then tried with a F_AMIP compset. However, then, during compilation the following message is generated:

> "File is missing ${DIN_LOC_ROOT}/atm/cam/sst/sst_HadOIBl_bc_48x96_1850_2012_c130411.nc
Invoke the following command to obtain them
./check_input_data -inputdata ${DIN_LOC_ROOT} -export"

However, when I try "./check_input_data -inputdata ${DIN_LOC_ROOT} -export", the following is output:

> "svn: URL 'https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata/atm/cam/sst/sst_HadOIBl_bc_48x96_1850_2012_c130411.nc' non-existent in that revision
File was not found in svn repo: https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata/atm/cam/sst/sst_HadOIBl_bc_48x96_1850_2012_c130411.nc"

I have checked and the file is indeed absent. However, I do not understand why it is being asked for, because when I checked the file ${CCSM_ROOT}/models/atm/cam/bld/namelist_files/namelist_defaults_cam.xml, and no such file ("sst_HadOIBl_bc_48x96_1850_2012_c130411.nc") is listed. The most similar file is listed under the entry:

"bndtvs hgrid="48x96" sim_year="1850-2000">atm/cam/sst/sst_HadOIBl_bc_48x96_1850_2008_c100128.nc"

Where else might there be a reference to the file "sst_HadOIBl_bc_48x96_1850_2012_c130411.nc" as being required? Or could someone perhaps suggest an alternative solution to this difficulty?

Thank you in advance

 

[eaton]

I see that the file sst_HadOIBl_bc_48x96_1850_2013_c140701.nc is in the inputdata repository, so that's the one I'd use.Note that the values for the SST datasets in CAM's namelist_defaults_cam.xml file are only used by CAM standalone scripts.  The CESM scripts put this information in the compset definition file.  I think changing this dataset can probably be done via the DOCN namelist or streams input though I don't know how to do that off the top of my head.  Another way would be to edit the file scripts/ccsm_utils/Case.template/config_compsets.xml.  Just look for the filename of the missing file and replace by the one above. 

 

[eaton]

I see that the file sst_HadOIBl_bc_48x96_1850_2013_c140701.nc is in the inputdata repository, so that's the one I'd use.Note that the values for the SST datasets in CAM's namelist_defaults_cam.xml file are only used by CAM standalone scripts.  The CESM scripts put this information in the compset definition file.  I think changing this dataset can probably be done via the DOCN namelist or streams input though I don't know how to do that off the top of my head.  Another way would be to edit the file scripts/ccsm_utils/Case.template/config_compsets.xml.  Just look for the filename of the missing file and replace by the one above. 

 

[cnrwil004@myuct_ac_za]

Thank you very much for the reply, Eaton. I could not find a reference to the file in the DOCN namelist, but the second suggestion worked easily. However, I now get the original error (regarding the cice component). Does this imply that it would not be possible to run a CESM F compset with the "eul" dycore? Or is there a different grid on which I should still try to get it to work?
Thanks again

 

[cnrwil004@myuct_ac_za]

Thank you very much for the reply, Eaton. I could not find a reference to the file in the DOCN namelist, but the second suggestion worked easily. However, I now get the original error (regarding the cice component). Does this imply that it would not be possible to run a CESM F compset with the "eul" dycore? Or is there a different grid on which I should still try to get it to work?
Thanks again

 

[eaton]

You should be able to run an F_AMIP compset at resolution T31_T31.  I tried (after fixing the name of the SST dataset) and encountered the same error you report.I'm able to run this configuration using a CAM standalone script by making use of the 1850-2005_cam4 use case (the CESM scripts are also using this) and by customizing the CAM namelist to use the new SST dataset, i.e., set the bndtvs and stream_year_last namelist variables appropriately.The problem with the CESM configuration is that the ice_in namelist file contains two references to '.../inputdata/share/domains/UNSET'.  The standalone build is providing references to '.../inputdata/atm/cam/ocnfrac/domain.camocn.48x96_gx3v7_100114.nc'.  The problem seems to be in the CESM scripts since I see the correct filename in the config_compsets.xml file, but somehow it's not getting into the ice_in file.  

 

[eaton]

You should be able to run an F_AMIP compset at resolution T31_T31.  I tried (after fixing the name of the SST dataset) and encountered the same error you report.I'm able to run this configuration using a CAM standalone script by making use of the 1850-2005_cam4 use case (the CESM scripts are also using this) and by customizing the CAM namelist to use the new SST dataset, i.e., set the bndtvs and stream_year_last namelist variables appropriately.The problem with the CESM configuration is that the ice_in namelist file contains two references to '.../inputdata/share/domains/UNSET'.  The standalone build is providing references to '.../inputdata/atm/cam/ocnfrac/domain.camocn.48x96_gx3v7_100114.nc'.  The problem seems to be in the CESM scripts since I see the correct filename in the config_compsets.xml file, but somehow it's not getting into the ice_in file.